Amylostatin J

Directly coupled LC-MS and LC-NMR were applied to identify and structurally characterize an acarbose degradation product A in acidic media. A comparative analysis of the stop-flow LC-NMR (1H and TOCSY) and LC-MS data provided evidence that A is structurally related to acarbose, differing from the parent compound in a number of subunits present in the molecule. Spectral analysis revealed that A was the alpha-glucosidase inhibitor amylostatin XG. Complementary information obtained from the two methods led to the structural elucidation of A which was later corroborated by high-resolution NMR spectroscopy of the isolated molecule.

A stereo chemical refinement of the crystalline complex between porcine pancreatic alpha-amylase and a pseudopentasaccharide from the amylostatin family has been performed through molecular mechanics calculations, using a set of parameters appropriate for protein and protein-carbohydrate interactions. The refinement provided a starting point for docking a maltopentaose moiety within the catalytic site, in the absence of water. A thorough exploration of the different orientations and conformations of maltopentaose established the sense of binding of the amylosic substrate in the amylase cleft. After optimising the geometry of the binding site, the conformations adopted by the four contiguous linkages could be rationalised by considering the environment, either hydrophobic or hydrophilic, of the different glucose moieties. Seemingly, details of the non-bonded interactions (hydrogen bonds, van der Waals and stacking interactions) that underlie this molecular recognition have been established. In particular, it was confirmed that the three acidic amino acids of the catalytic site (Asp197, Asp300 and Glu233) are close to their glucosidic target, and that there is no steric reason to propose an alteration of the 4C1 conformation of the glucose residue prior to hydrolysis. However, in the absence of water molecules, it is difficult to elucidate the details of the catalysis. Additional macroscopic information has been gained, such as the impossibility to fit a double-helical arrangement of amylose chains in the amylasic cleft. This explains why some native starches containing such motifs resist amylolytic enzymes. Tentative models involving longer amylosic chains have been elaborated, which extend our knowledge of the interaction and orientation of starch fragments in the vicinity of the hydrolytic sites.